The true OAV, calibrated to the swaption market, looks very different from the static, zero-volatility valuation, doesn't it? The graph actually portrays a picture of how wrong the IO valuation multiple and the Greeks can be if volatility is ignored or even mis-modeled. For example, using a constant volatility number (such as the one on 3-yr into 10-yr swaption) is better than nothing, but it will likely fall short in accuracy once valuation for a wide range of coupons and risk scenarios are required.

OAS or OAV?
Our main point - to consider market volatility when valuing MSRs - should not be confused with the idea of pricing them under constant option-adjusted spread (OAS). A constant-OAS version of OAV is widely used today in the mortgage market, but it may be insufficiently accurate in measuring market risk of MSRs. Since the OAS is paid for bearing a non-market risk (prepayment surprises, for example), one should naturally expect it to be higher for IOs and MSRs stripped off premium pools and lower for discount pools. This is the way liquid IOs securities are traded today. Since IOs and MSRs are close cousins, we recommend using a variable-OAS pricing model for MSRs.

Indirect volatility effect
Above, we discussed the "direct" effect (all else fixed) of volatility in assessing values of IOs and MSRs. When volatility rises, we expect mortgage indices (such as MTGEFNCL) to rise as well. This should happen because MBS investors will seek a higher compensation for being short an option. Under these circumstances, according to most prepayment models, prepay speeds should slow down, thereby boosting the values of IOs. This "indirect" volatility effect can sometimes result in a counter-intuitive value change. For example, a FAS 133 risk manager should be aware of volatility change that occurred between two sequential measurements. Of course, when a relevant mortgage index is directly employed in risk measurement and hedging as a separate factor, it will ultimately absorb the indirect volatility effect.